LAB NOTE · OPERATING PRINCIPLE
Public Narrative, Private Control
A personal AI OS can be public enough to explain itself without being public enough to operate itself.
The principle
Publish narrative, architecture, project maps, coarse health, and selected lessons. Keep secrets, production controls, raw logs, quotas, admin dashboards, payment flows, and private agent channels outside the public website.
Why it matters
AI systems become more useful when they can be explained, audited, and referenced. They also become more fragile when public pages accidentally become control surfaces. The right public layer helps collaborators understand the system while preserving operational safety.
Practical rule
Default answer
If a piece of information helps explain the system, publish it. If it can operate, compromise, identify, quota-drain, or mutate the system, keep it private or behind strong authentication.